What Is a Statistically Adequate Model and Why Is It Important?
For an ordinary least squares (OLS)-type regression, the researcher assumes that the errors are normally, identically, and independently distributed (NIID). In practice, it is the residuals (the estimated errors) that must satisfy these assumptions in order for the researcher to draw valid inference from the results of their regression. In other words, if the residuals are not NIID in the finite sample, the model is considered to be misspecified, and at least some of the inference drawn from the results cannot be trusted. Sometimes it is the standard errors of the coefficients that are biased, and sometimes it is the coefficients themselves, or both. Either way, any conclusions ...
Get Building Better Econometric Models Using Cross Section and Panel Data now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.